Exercise 1 - Poisson Image completion using Discrete Poisson Linear Systems.

Exercise 1 - Poisson• Image completion

using Discrete Poisson Linear Systems

Exercise 1 - Poissonif Ix,y is not black Px,y = Ix,y

else 4Px,y - Px-1,y - Px+1,y - Px,y-1 - Px,y+1 = 0

(m)x(n) image(m·n)x(m·n) linear system

Resolution: 480x320Linear System: 153600x 153600

Resolution: 576x300Linear System: 172800x172800

Exercise 1 - PoissonImageCloning

Representing an imageUsing a 1D array to represent a 2D image.index(x,y) = y*numOfColumns + x. This basically serializes the image row after row.

Coordinate (0,0) is always the upper left corner.The y coordinates increment going down, as opposed to normal Cartesian coordinates.

Representing ColorBufferedImage img = ...int c = img.getRGB(x, y);// c == 16597993 ???A pixel is represented by 3 values for the 3 color channels.Each color value is a number in the range [0,255].Fits a single unsigned byte.To extract the separate values, use bit operations

16597993 = 0x00FD43E9

blue E9=233 green 43=67 red FD=253